Persuasion by Shaping Beliefs about Multidimensional Features of a Thing

Kazunori Terada, Yasuo Noma, Masanori Hattori

Abstract: Research has demonstrated the effectiveness of personalization in persuasive agents, recommendation agents, and nudge agents. Ultimate personalization targets the presentation of information tailored to an individual's nuanced beliefs and utilities, rather than relying on broad attributes such as personality traits, age, or gender. Multi-attribute utility theory suggests that the utility of a thing is determined by the sum of the utilities given to its various features. In our research, we developed a method to enhance the personal utility of a thing by addressing and manipulating people's beliefs about the features of a thing. We conducted an experiment (n=197) to verify whether the proposed method can increase the participants' utility of a fully autonomous vehicle, as a target of persuasion. Among 13 propositions (features) that constitute the concept of fully autonomous vehicles, in a semi-structured dialog, a virtual agent presented counter-propositions to the top propositions that each participant assigned the most negative utilities. Before and after the dialog, the monetary value of fully autonomous vehicles, the desire to ride them, and the social obligation to accept them were measured. The results showed that the proposed method improved the social obligation to accept fully autonomous vehicles more than the baseline method and the non-personalized method, but had no effect on the monetary value and the desire to ride them. This suggests that personalized belief manipulation may not be effective in enhancing the "want to" desire or utility of a thing, but may only improve the thought of "ought to do".

External IDs:dblp:conf/atal/TeradaNH24